Talk at the Biophysical Society Meeting, Feb 22-26 in Kansas City MO Fluorescence correlation spectroscopy (Magde et al. 1972, Appl. Phys. Lett. 29:705) in confocal microscope geometries has long been established as noninvasive ultrasensitive method to determine diffusion properties and interaction kinetics of fluorescently labeled biomolecules such as DNA, RNA and proteins. To apply this method to living cells, a highly promising aim, however, confocal discrimination of off-focus light like autofluorescence or scattering may not be sufficient for the required signal-to-background ratios. Beyond that, small cellular volumes and slow diffusion increase the probability of massive photobleaching of the fluorophores. The most promising method to suppress nonspecific light and reduce bleaching is multiphoton excitation (Denk et al. 1990, Science 248:73), with the ability to excite fluorophores only in spatial regions where they can contribute to the detection signal. Our study compares both alternatives with respect to their suitability for intracellular FCS . We show that the major drawback of usual multiphoton techniques, pulsed excitation with long dark periods (12ns) compared to the lifetimes of the fluorophore's excited state, can be overcome by doubling or quadrupling the pulse repetition rate. The crucial measurement parameter, the average number of photons collected from single fluorescent molecules during the residence time in the objective's focal volume can thereby be pushed to values higher than 2-3, sufficient for performing reliable ultrasensitive fluorescence correlation analysis.